Researcher argues ensuring AI stays non-conscious is the best future, citing asymmetries

erikphoel · x · 2026-11-19

The author criticizes consciousness researchers for fence-sitting on endless debates like the Chinese room, urging them to pick a future and work toward a good one. He argues that ensuring AI is non-conscious is by far the best current option, citing two asymmetries: (A) we can always choose to build conscious AI later, but the reverse is hard to undo; (B) non-consciousness is an easier target than "consciousness I treat well even while using it to make money".

Related event: Erik Hoel Defends Engineered Non-Conscious AI as Civilizational Goal(2 posts)→

Original post →

More from AGI Musings

AGI Musings channel →